🏋️ Challenges — Module 04 (Automation)

1. Headless: run opencode run --format json "List the TODOs" and parse it with jq.
2. Sessions: chain two commands with --continue, then export the session with --sanitize.
3. CI: copy examples/ci/opencode-review.yml into a test repo, set the secret and open a PR with an intentional bug.
4. Quality gate: make the job fail when the output contains a high-severity finding.

💡 Solution (challenge 1)

opencode run --format json "List the TODO comments" | jq -r '.text'

💡 Solution (challenge 3 — essence)

- run: npm install --global @opencode/cli
- run: opencode run --standalone --agent reviewer "Review the PR and list findings by severity."
  env:
    ANTHROPIC_API_KEY: ${{ secrets.ANTHROPIC_API_KEY }}

➡️ Module 05